its not braking anymore

main.py

@@ -23,6 +23,7 @@ from envs.econ_wrapper import EconVecEnv
 from stable_baselines3.common.callbacks import BaseCallback
 import yaml
 import time
+from threading import Thread 
 
 env_config = {
     # ===== SCENARIO CLASS =====
@@ -61,7 +62,7 @@ env_config = {
     'allow_observation_scaling': True,
     'dense_log_frequency': 100, 
     'world_dense_log_frequency':1,
-    'energy_cost':0.21,
+    'energy_cost':0,
     'energy_warmup_method': "auto",
     'energy_warmup_constant': 4000,
     
@@ -116,7 +117,7 @@ eval_env_config = {
     'allow_observation_scaling': True,
     'dense_log_frequency': 10, 
     'world_dense_log_frequency':1, 
-    'energy_cost':0.21,
+    'energy_cost':0,
     'energy_warmup_method': "auto",
     'energy_warmup_constant': 4000,
     
@@ -223,23 +224,33 @@ baseEconWrapper=BaseEconWrapper(econ)
 baseEconWrapper.run()
 mobileRecieverEconWrapper=RecieverEconWrapper(base_econ=baseEconWrapper,agent_classname="BasicMobileAgent")
 tradeRecieverEconWrapper=RecieverEconWrapper(base_econ=baseEconWrapper,agent_classname="TradingAgent")
+sb3_traderConverter=SB3EconConverter(tradeRecieverEconWrapper,econ,"TradingAgent")
 sb3Converter=SB3EconConverter(mobileRecieverEconWrapper,econ,"BasicMobileAgent")
 #obs=sb3Converter.reset()
 #vecenv=EconVecEnv(env_config=env_config)
 
 monenv=VecMonitor(venv=sb3Converter,info_keywords=["social/productivity","trend/productivity"])
-
+montraidingenv=VecMonitor(venv=sb3_traderConverter)
 #normenv=VecNormalize(sb3Converter,norm_reward=False,clip_obs=1)
 #stackenv=vec_frame_stack.VecFrameStack(venv=monenv,n_stack=10)
 obs=monenv.reset()
 
 
+# define training functions
+def train(model,timesteps, econ_call,process_bar,name,db,index):
+    db[index]=model.learn(total_timesteps=timesteps,progress_bar=process_bar,reset_num_timesteps=False,tb_log_name=name,callback=TensorboardCallback(econ_call))
 
 
 
-runname="run_{}".format(int(np.random.rand()*100))
+# prepare training
+run_number=int(np.random.rand()*100)
+runname="run_{}".format(run_number)
+model_db=[] # object for storing model
+
 
 model = MaskablePPO("MlpPolicy",n_steps=int(env_config['episode_length']*2),ent_coef=0.1, vf_coef=0.8 ,gamma=0.95, learning_rate=5e-3,env=monenv, seed=225,verbose=1,device="cuda",tensorboard_log="./log")
+model_trade=MaskablePPO("MlpPolicy",n_steps=int(env_config['episode_length']*2),ent_coef=0.1, vf_coef=0.8 ,gamma=0.95, learning_rate=5e-3,env=montraidingenv, seed=225,verbose=1,device="cuda",tensorboard_log="./log")
+
 n_agents=econ.n_agents
 total_required_for_episode=n_agents*env_config['episode_length']
 print("this is run {}".format(runname))
@@ -249,11 +260,21 @@ eval_base_econ.run()
 eval_mobileRecieverEconWrapper=RecieverEconWrapper(eval_base_econ,"BasicMobileAgent")
 time.sleep(0.5)
 eval_sb3_converter=SB3EconConverter(eval_mobileRecieverEconWrapper,eval_econ,"BasicMobileAgent")
+
 while True:
     # Create Eval ENV
     vec_mon_eval=VecMonitor(venv=eval_sb3_converter)
     #Train
-    model=model.learn(total_timesteps=total_required_for_episode*10,progress_bar=True,reset_num_timesteps=False,tb_log_name=runname,callback=TensorboardCallback(econ=econ))
+    runname="run_{}_{}".format(run_number,"basic")
+
+    thread_model=Thread(target=train,args=(model,total_required_for_episode*10,econ,True,runname,model_db,0))
+    runname="run_{}_{}".format(run_number,"trader")
+    thread_model_traid=Thread(target=train,args=(model_trade,total_required_for_episode*10,econ,False,runname,model_db,1))
+    
+    thread_model.start()
+    thread_model_traid.start()
+    thread_model.join()
+    thread_model_traid.join()
     #normenv.save("temp-normalizer.ai")